Views
No views yet
UIGEN-T3-4B-Preview is Tesslate’s advanced UI generation model built on the Qwen3 architecture, designed to produce both individual components and full web pages using a hybrid reasoning system that offers modes for guided reasoning as well as fast, direct code generation. Trained with the UIGenEval benchmark for robust frontend development, it supports clean HTML, CSS, responsive layouts, and technical quality, while also providing design tokens with each output for easy site-wide customization. Released under a research-only, non-commercial license, UIGEN-T3-4B-Preview is tailored for rapid prototyping, internal tooling, and component library creation, requiring at least a 4GB VRAM GPU for inference and compatible with popular ML libraries, but not intended for production use or commercial deployment without special licensing
| File name | Size | Quant Type |
|---|---|---|
| UIGEN-T3-4B-Preview.F32.gguf | 16.1 GB | F32 |
| UIGEN-T3-4B-Preview.BF16.gguf | 8.05 GB | BF16 |
| UIGEN-T3-4B-Preview.F16.gguf | 8.05 GB | F16 |
| UIGEN-T3-4B-Preview.Q8_0.gguf | 4.28 GB | Q8_0 |
| UIGEN-T3-4B-Preview.Q6_K.gguf | 3.31 GB | Q6_K |
| UIGEN-T3-4B-Preview.Q5_K_M.gguf | 2.89 GB | Q5_K_M |
| UIGEN-T3-4B-Preview.Q5_K_S.gguf | 2.82 GB | Q5_K_S |
| UIGEN-T3-4B-Preview.Q4_K_M.gguf | 2.5 GB | Q4_K_M |
| UIGEN-T3-4B-Preview.Q4_K_S.gguf | 2.38 GB | Q4_K_S |
| UIGEN-T3-4B-Preview.Q3_K_L.gguf | 2.24 GB | Q3_K_L |
| UIGEN-T3-4B-Preview.Q3_K_M.gguf | 2.08 GB | Q3_K_M |
| UIGEN-T3-4B-Preview.Q3_K_S.gguf | 1.89 GB | Q3_K_S |
| UIGEN-T3-4B-Preview.Q2_K.gguf | 1.67 GB | Q2_K |
